Computer simulations of an iron-powder flame produced sharply different answers to a basic engineering question: how hot must the surrounding gas be before particles ignite? In the oxide-layer model, the minimum simulated ignition temperature was 1,125 kelvin, while complete oxidation was reached at 1,250 kelvin. A first-order model put those points at 800 kelvin and 900 kelvin. Both sets differed from the experimental results used as a reference.
The study asked how oxidation-model choice, flame enclosure and radiation treatment affect simulated ignition and oxidation in a jet-in-hot-coflow burner. It also compared the simulated flame at different coflow temperatures.
A flame that develops in stages
At a coflow temperature of 1,200 kelvin, ignition appeared around the circumference of the particle jet, showing a localized pattern. Oxygen depletion marked ignition onset at roughly 100 millimetres above the injection nozzle. The coflow and particle jet broke up at roughly 200 millimetres.
The temperature comparison changed that pattern. No ignition was observed at 1,100 kelvin. Ignition appeared at 1,200 and 1,300 kelvin, and the highest-temperature case showed more uniform ignition and earlier completion of oxidation.
Two chemistry descriptions, two sets of thresholds
The researchers used large-eddy simulation, or LES, to calculate the flow and Lagrangian point-particles to track particles through it. The cases compared open and enclosed flames, first-order and oxide-layer kinetics, and simplified Stefan-Boltzmann and P1 radiation treatments.
The input was a polydisperse powder, meaning it contained several particle sizes. The reported bins were 25, 35, 45, 55 and 65 micrometres. The D3,2 size measure was 46.93 micrometres, and the total injection rate was 300 milligrams per second.
The disagreement continued when the simulations were compared with the referenced experiment. The conclusion says the first-order model overpredicted oxidation, while the oxide-layer model underpredicted it. For the oxide-layer calculation, ignition was approximately 1,125 kelvin and complete oxidation approximately 1,250 kelvin, versus experimental values of 950 kelvin and 1,050 kelvin.
The simulated setup mattered
In the comparison of flame configurations, enclosed or wall-bounded flames had a higher oxidation degree than open flames. The P1 radiation case had slightly lower oxidation than the simplified model, although the paper described that radiation difference as insignificant.
One especially clear pattern appeared in a wall-bounded case at 1,150 kelvin using simplified radiation. Its overall oxidation degree was 0.805. Farther downstream, particles were predominantly either unignited, with an oxidation degree below 0.2, or completely oxidized, with a value close to 1. In this analyzed case, the global oxidation degree reflected ignition probability.
Particle size added another, smaller difference. The 45-micrometre bin contained the largest number of unburnt particles because that bin was more numerous. After accounting for the number of particles in each size bin, ignition failure was slightly more prevalent among larger particles.
Why the numbers need caution
The paper flags several unresolved inputs and assumptions as relevant to the simulation-experiment discrepancy: particle-injection temperature and morphology, missing detailed jet-in-hot-coflow injection boundary conditions, and the need for an oxidation model for irregular, rough-surface particles.
The chemical scope was limited. The carrier phase included only oxygen and nitrogen, with no gas-phase reactions, while particle chemistry was represented as a single-stage iron-to-iron-oxide reaction.
This is an arXiv version 1 preprint dated 28 August 2026. The authors identify detailed injection conditions and an oxidation model for irregular, rough-surface particles as priorities for resolving the reported mismatch.
Paper data and sources
Original title: LES of iron-powder combustion in a jet-in-hot-coflow burner - Insights on flame structure and ignition characteristics
Authors: Shyam Hemamalini, XiaoCheng Mi
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-28
DOI: Not available
Original paper · Full text